Evaluation of blood glucose level control in type 1 diabetic patients using deep reinforcement learning.

Diabetes mellitus is a disease associated with abnormally high levels of blood glucose due to a lack of insulin. Combining an insulin pump and continuous glucose monitor with a control algorithm to deliver insulin is an alternative to patient self-management of insulin doses to control blood glucose levels in diabetes mellitus patients. In this work, we propose a closed-loop control for blood glucose levels based on deep reinforcement learning. We describe the initial evaluation of several alternatives conducted on a realistic simulator of the glucoregulatory system and propose a particular implementation strategy based on reducing the frequency of the observations and rewards passed to the agent, and using a simple reward function. We train agents with that strategy for three groups of patient classes, evaluate and compare it with alternative control baselines. Our results show that our method is able to outperform baselines as well as similar recent proposals, by achieving longer periods of safe glycemic state and low risk.

Enhancing Extensive and Remote LoRa Deployments through MEC-Powered Drone Gateways.

The distribution of Internet of Things (IoT) devices in remote areas and the need for network resilience in such deployments is increasingly important in smart spaces covering scenarios, such as agriculture, forest, coast preservation, and connectivity survival against disasters. Although Low-Power Wide Area Network (LPWAN) technologies, like LoRa, support high connectivity ranges, communication paths can suffer from obstruction due to orography or buildings, and large areas are still difficult to cover with wired gateways, due to the lack of network or power infrastructure. The proposal presented herein proposes to mount LPWAN gateways in drones in order to generate airborne network segments providing enhanced connectivity to sensor nodes wherever needed. Our LoRa-drone gateways can be used either to collect data and then report them to the back-office directly, or store-carry-and-forward data until a proper communication link with the infrastructure network is available. The proposed architecture relies on Multi-Access Edge Computing (MEC) capabilities to host a virtualization platform on-board the drone, aiming at providing an intermediate processing layer that runs Virtualized Networking Functions (VNF). This way, both preprocessing or intelligent analytics can be locally performed, saving communications and memory resources. The contribution includes a system architecture that has been successfully validated through experimentation with a real test-bed and comprehensively evaluated through computer simulation. The results show significant communication improvements employing LoRa-drone gateways when compared to traditional fixed LoRa deployments in terms of link availability and covered areas, especially in vast monitored extensions, or at points with difficult access, such as rugged zones.

Impact of SCHC Compression and Fragmentation in LPWAN: A Case Study with LoRaWAN.

The dawn of the Internet of Things (IoT) paradigm has brought about a series of novel services never imagined until recently. However, certain deployments such as those employing Low-Power Wide-Area Network (LPWAN)-based technologies may present severe network restrictions in terms of throughput and supported packet length. This situation prompts the isolation of LPWAN systems on islands with limited interoperability with the Internet. For that reason, the IETF's LPWAN working group has proposed a Static Context Header Compression (SCHC) scheme that permits compression and fragmentation of and IPv6/UDP/CoAP packets with the aim of making them suitable for transmission over the restricted links of LPWANs. Given the impact that such a solution can have in many IoT scenarios, this paper addresses its real evaluation in terms not only of latency and delivery ratio improvements, as a consequence of different compression and fragmentation levels, but also of the overhead in end node resources and useful payload sent per fragment. This has been carried out with the implementation of middleware and using a real testbed implementation of a LoRaWAN-to-IPv6 architecture together with a publish/subscribe broker for CoAP. The attained results show the advantages of SCHC, and sustain discussion regarding the impact of different SCHC and LoRaWAN configurations on the performance. It is highlighted that necessary end node resources are low as compared to the benefit of delivering long IPv6 packets over the LPWAN links. In turn, fragmentation can impose a lack of efficiency in terms of data and energy and, hence, a cross-layer solution is needed in order to obtain the best throughput of the network.

Empowering the Internet of Vehicles with Multi-RAT 5G Network Slicing.

Internet of Vehicles (IoV) is a hot research niche exploiting the synergy between Cooperative Intelligent Transportation Systems (C-ITS) and the Internet of Things (IoT), which can greatly benefit of the upcoming development of 5G technologies. The variety of end-devices, applications, and Radio Access Technologies (RATs) in IoV calls for new networking schemes that assure the Quality of Service (QoS) demanded by the users. To this end, network slicing techniques enable traffic differentiation with the aim of ensuring flow isolation, resource assignment, and network scalability. This work fills the gap of 5G network slicing for IoV and validates it in a realistic vehicular scenario. It offers an accurate bandwidth control with a full flow-isolation, which is essential for vehicular critical systems. The development is based on a distributed Multi-Access Edge Computing (MEC) architecture, which provides flexibility for the dynamic placement of the Virtualized Network Functions (VNFs) in charge of managing network traffic. The solution is able to integrate heterogeneous radio technologies such as cellular networks and specific IoT communications with potential in the vehicular sector, creating isolated network slices without risking the Core Network (CN) scalability. The validation results demonstrate the framework capabilities of short and predictable slice-creation time, performance/QoS assurance and service scalability of up to one million connected devices.

LPWAN-Based Vehicular Monitoring Platform with a Generic IP Network Interface.

Remote vehicle monitoring is a field that has recently attracted the attention of both academia and industry. With the dawn of the Internet of Things (IoT) paradigm, the possibilities for performing this task have multiplied, due to the emergence of low-cost and multi-purpose monitoring devices and the evolution of wireless transmission technologies. Low Power-Wide Area Network (LPWAN) encompasses a set of IoT communication technologies that are gaining momentum, due to their highly valued features regarding transmission distance and end-device energy consumption. For that reason, in this work we present a vehicular monitoring platform enabled by LPWAN-based technology, namely Long Range Wide Area Network (LoRaWAN). Concretely, we explore the end-to-end architecture considering vehicle data retrieving by using an On-Board Diagnostics II (OBD-II) interface, their compression with a novel IETF compression scheme in order to transmit them over the constrained LoRaWAN link, and information visualization through a data server hosted in the cloud, by means of a web-based dashboard. A key advance of the proposal is the design and development of a UNIX-based network interface for LPWAN communications. The whole system has been tested in a university campus environment, showing its capabilities to remotely track vehicle status in real-time. The conducted performance evaluation also shows high levels of reliability in the transmission link, with packet delivery ratios over 95%. The platform boosts the process of monitoring vehicles, enabling a variety of services such as mechanical failure prediction and detection, fleet management, and traffic monitoring, and is extensible to light vehicles with severe power constraints.

Enhancing LoRaWAN Security through a Lightweight and Authenticated Key Management Approach.

Luckily, new communication technologies and protocols are nowadays designed considering security issues. A clear example of this can be found in the Internet of Things (IoT) field, a quite recent area where communication technologies such as ZigBee or IPv6 over Low power Wireless Personal Area Networks (6LoWPAN) already include security features to guarantee authentication, confidentiality and integrity. More recent technologies are Low-Power Wide-Area Networks (LP-WAN), which also consider security, but present initial approaches that can be further improved. An example of this can be found in Long Range (LoRa) and its layer-two supporter LoRa Wide Area Network (LoRaWAN), which include a security scheme based on pre-shared cryptographic material lacking flexibility when a key update is necessary. Because of this, in this work, we evaluate the security vulnerabilities of LoRaWAN in the area of key management and propose different alternative schemes. Concretely, the application of an approach based on the recently specified Ephemeral Diffie⁻Hellman Over COSE (EDHOC) is found as a convenient solution, given its flexibility in the update of session keys, its low computational cost and the limited message exchanges needed. A comparative conceptual analysis considering the overhead of different security schemes for LoRaWAN is carried out in order to evaluate their benefits in the challenging area of LP-WAN.

Relación del factor geológico en la tasa de sedimentación de Embalses Sierras de Córdoba y San Luis

En este trabajo los autores analizan la relación existente entre la naturaleza geológica y los procesos geomorfológicos de índole dinámico y la tasa de sedimentación en embalses. Como representativos de la provincia geológica que los incluye y en base a las características de sus cuencas de aporte, se seleccionaron los embalses San Roque en la Pcia. de Córdoba y La Florida y Cruz de Piedra en la Pcia. de San Luis

Red telemétrica cuenca del río San Antonio: provincia de Córdoba

En la cuenca del río San Antonio, de aproximadamente 540 km2., se producen durante el periodo estival tormentas de tipo convectivo que generan en un lapso de 7 a 8 horas importantes crecidas. En la citada cuenca se detecta, en este sentido, escasez de información básica histórica, dificultando la planificación, aprovechamiento y control del recurso hídrico. Los problemas mencionados de obtención básica, y la formulación de técnicas de pronóstico hidrológico, encuentran solución en equipamientos electrónicos que posibilitan la realización de estudios detallados suministrando una herramienta eficaz para el conocimiento del recurso hídrico de la cuenca disponiendo de datos a tiempo real para su ingreso a un modelo matemático que permita predecir caudales y crecidas, conociendo la distribución espacial y temporal de la precipitación en la cuenca. Frente a esta situación, el INCYTH, a través del CIHRSA, instaló 11 estaciones remotas de medición. Con el equipamiento mencionado (en funciones desde 1986) se implementó para los periodos estivales 87/88 y 88/89 un sistema de alerta de crecidas para la cuenca del río San Antonio a tiempo real

Red telemétrica cuenca del río San Antonio: provincia de Córdoba

En la cuenca del río San Antonio, de aproximadamente 540 km2., se producen durante el periodo estival tormentas de tipo convectivo que generan en un lapso de 7 a 8 horas importantes crecidas. En la citada cuenca se detecta, en este sentido, escasez de información básica histórica, dificultando la planificación, aprovechamiento y control del recurso hídrico. Los problemas mencionados de obtención básica, y la formulación de técnicas de pronóstico hidrológico, encuentran solución en equipamientos electrónicos que posibilitan la realización de estudios detallados suministrando una herramienta eficaz para el conocimiento del recurso hídrico de la cuenca disponiendo de datos a tiempo real para su ingreso a un modelo matemático que permita predecir caudales y crecidas, conociendo la distribución espacial y temporal de la precipitación en la cuenca. Frente a esta situación, el INCYTH, a través del CIHRSA, instaló 11 estaciones remotas de medición. Con el equipamiento mencionado (en funciones desde 1986) se implementó para los periodos estivales 87/88 y 88/89 un sistema de alerta de crecidas para la cuenca del río San Antonio a tiempo real

Estudio batimétrico en cinco embalses de la provincia de San Luis

Tiene por objetivo conocer el atarquinamiento y en consecuencia la capacidad útil actual que presentan cinco embalses de la provincia de San Luis, a saber: Dique Cruz de Piedra, La Florida, Potrero de los Funes, San Felipe y Luján. A tal efecto, se relevaron por medio de batimetría los embalses mencionados, y se ubicó para cada uno de ellos el volúmen actual. Se determinó además, por comparación con el volumen primitivo, el volumen de sedimentos acumulado. Los resultados indican, por un lado, una sedimentación elevada en los embalses Cruz de Pidera y Luján, y un proceso normal de colmatación en los restantes